A Multi-task Learning Approach for Mandarin-English Code-Switching Conversational Speech Recognition.
Xiao SongYi LiuDaming YangYuexian ZouPublished in: ISICA (1) (2017)
Keyphrases
- speech recognition
- multi task learning
- speech recognition technology
- multi task
- automatic speech recognition
- learning tasks
- gaussian processes
- natural language
- speaker independent
- broadcast news
- hidden markov models
- high order
- pattern recognition
- conversational speech
- learning problems
- language model
- speech recognizer
- multi modal
- spoken language
- speech signal
- theoretical analysis
- noisy environments
- speech synthesis
- learning models
- speaker identification
- transfer learning
- speech recognition systems
- learning algorithm
- cross language
- spoken document retrieval
- machine learning
- kernel methods
- text mining
- training data
- decision trees